Research exploring the feasibility of deriving population estimates from remotely sensed data demonstrates that objects in the urban landscape can be identified and incorporated into a population estimates system based on the housing unit method. Nonetheless, this research also reveals shortcomings in the technology producing the input files used in the automatic detection of objects. The problem involves the assumption and techniques used when converting high resolution images into digital elevation models (DEM). DEM files serve as input to the programs used in the detection of housing units. Efforts to correctly identify housing units are time-consuming and error-prone without clear and distinct DEMs. The objectives of this Phase I SBIR application are to further refine strengthen and test the software employed in transforming satellite and aerial imagery into digital elevation model (DEMs). DEM files serve as input to the programs used in the detection of housing units. Efforts to correctly identify housing units are time-consuming and error-prone without clear and distinct DEMs. The objectives of this Phase I SBIR application are to further refine, strengthen and test the software employed in transforming satellite and aerial imagery into digital elevation models. Specific goals of this Phase I proposal include: 1) modifying and coding new assumptions into the DEM software, 2) testing the accuracy and reliability of the digital elevation code on new sub1, aerial imagery, and 3) designing a new GUI for use in the pre- processing phase of DEM building. PROPOSED COMMERCIAL APPLICATIONS: The commercial value of this specific research is best understood when viewed as part of a larger effort to produce an automated system for deriving population estimates of user defined areas based on current, remotely sensed data. Such a system will serve a wide range of commercial interests seeking "up-to-the-minute" counts and measures of population and housing change.